Guide
EU AI Act high-risk AI systems
Updated June 16, 2026 · By Max Langley, AI Audits EU
An AI system is high-risk under the EU AI Act in two situations: when it is a safety component of a product already regulated under the EU laws in Annex I, or when its use falls into one of the categories in Annex III. High-risk systems carry the Act's heaviest obligations, so classification is the first thing to get right.
The two routes to high-risk
The first route is Annex I. If your AI is a safety component of a product covered by existing EU product law, such as machinery, medical devices, in-vitro diagnostics, lifts, or toys, and that product already requires third-party conformity assessment, the AI is high-risk. The second route is Annex III, a list of standalone use cases that are high-risk because of the impact they have on people's rights and access to opportunity.
The Annex III high-risk categories
- 1. Biometrics. Remote biometric identification, biometric categorisation, and emotion recognition, where permitted.
- 2. Critical infrastructure. Safety components in the management of utilities, traffic, and digital infrastructure.
- 3. Education and vocational training. Admissions, scoring, and monitoring of learners.
- 4. Employment and worker management. Recruitment, screening, ranking, promotion, and task allocation.
- 5. Essential private and public services. Creditworthiness and credit scoring, life and health insurance risk, and emergency service dispatch.
- 6. Law enforcement. Risk assessments, evidence evaluation, and profiling.
- 7. Migration, asylum, and border control. Risk assessments and application processing.
- 8. Administration of justice and democratic processes. Assisting judicial decisions and influencing elections.
The significant-risk filter
Being in an Annex III category does not always make a system high-risk. The Act carves out systems that do not pose a significant risk to health, safety, or fundamental rights, for example a tool that performs a narrow procedural task, improves the output of a finished human activity, or flags patterns without replacing human judgment. The catch is that any system which profiles people is always high-risk, and if you decide a system is out of scope you have to document the reasoning and register that conclusion. Treat the carve-out as the exception, not the default.
What high-risk triggers
A high-risk classification brings the full obligation set: a risk management system, data governance, technical documentation, automatic logging, transparency and instructions for use, human oversight, and accuracy, robustness, and cybersecurity. Before the system goes on the market the provider completes a conformity assessment, draws up an EU declaration of conformity, affixes CE marking, and registers the system in the EU database. After launch, post-market monitoring and serious-incident reporting apply.
What makes an AI system high-risk under the EU AI Act?
There are two routes. A system is high-risk if it is a safety component of a product already regulated under EU law listed in Annex I, such as machinery, medical devices, or toys, and that product needs third-party conformity assessment. It is also high-risk if it falls into one of the use cases listed in Annex III, such as employment, biometrics, or credit scoring.
What are the Annex III high-risk categories?
Annex III covers biometrics, critical infrastructure, education and vocational training, employment and worker management, access to essential private and public services such as credit and insurance, law enforcement, migration, asylum and border control, and the administration of justice and democratic processes.
Is every system in those categories automatically high-risk?
No. An Annex III system is not high-risk if it does not pose a significant risk to health, safety, or fundamental rights, for example when it performs a narrow procedural task or only improves the result of a completed human activity. But a system that profiles individuals is always treated as high-risk, and you must document any decision that a system is out of scope.
What obligations does a high-risk classification trigger?
High-risk providers must run a risk management system, meet data governance standards, prepare technical documentation, keep logs, ensure transparency and human oversight, and meet accuracy, robustness, and cybersecurity requirements. They must complete a conformity assessment, affix CE marking, and register the system in the EU database before placing it on the market.
Sources
- Regulation (EU) 2024/1689 (the AI Act), Articles 6 to 7 and Annexes I and III, EUR-Lex, eur-lex.europa.eu.
- European Commission, Regulatory framework on artificial intelligence, digital-strategy.ec.europa.eu.
Is your system high-risk?
Classification is the first and most consequential call. Tell us what your system does and we will assess its risk tier and what follows.
This guide is general information, not legal advice. Confirm your obligations with qualified counsel.